Re: a small query about Value numbering

"Jeremy Wright" <Jeremy.Wright@microfocus.com>
26 May 2006 12:48:37 -0400

          From comp.compilers
Related articles
a small query about Value numbering balu.ugandhar@gmail.com (yugandhar balaji krishna atukuri) (2006-05-26)
Re: a small query about Value numbering Jeremy.Wright@microfocus.com (Jeremy Wright) (2006-05-26)
| List of all articles for this month |
From: "Jeremy Wright" <Jeremy.Wright@microfocus.com>
Newsgroups: comp.compilers
Date: 26 May 2006 12:48:37 -0400
Organization: Compilers Central
References: 06-05-079
Keywords: analysis
Posted-Date: 26 May 2006 12:48:36 EDT

According to "Advanced Compiler Design and Implementation", Steven S.
Muchnick, pg343, in the following example Global Common SubExpression
and Partial Redundancy Elimination both eliminate the third "2 * i" but
Value Numbering does not.


        read(i)
        l := 2 * i
        if i > 0 goto L1
        j := 2 * i
        goto L2
L1: k := 2 * i
L2:


As others have commented before, Muchnick's book is excellent but there
is no hand holding. He expects you keep up with him.


> can anybody tell me a "case" which cannot be covered by VALUE
> NUMBERING and covered by Common subexpression elimination.
> personally i believe that Value numbering is Superior to common
> subexpression elimination and value numbering can optimize all the
> common subexpressions that can be optimized by common subexpression
> elimination.

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.